What do you guys think? a 40% IPC bump puts it in Haswell/Broadwell range. and with 8 Cores and hyperthreading.. do you think this is a good alternitve to Haswell-E or the New Broadwell=E? if they can get the price in check? like a normal AMD FX part?
they have yet to release more info than what is already rumored.
looks like its going to be LGA, 8-core for desktops, 14nm FinFET, it will be for enthusiasts and servers market.
sounds its going to be good enough to go against 1150/alike sockets, not sure about 2011lga.
2 Likes
im excited about the x480. i have a mostly i7 intel/gtx setup. but im tired of the super large gpus. looking fowards to build a good performance/cheaper build
1 Like
Sounds good, but is it true?
Can we really know this is true till we get the reviews out?
Cause the FuryX release was pretty hyped up too, just saying. I feel like if the numbers are true the market is about to shift.
FuryX got hyped up from AMD announcing it as HBM to the enthusiasts to "TITAN KILLER X"-nonsense.
I thought it performed better then Titan X? The issue was that 50% of the games worked faster on NVIDIA while the other %50 worked better on Fury X. The problem with Fury cards is the 4GB memory however! If they were 8GB I might have considered getting one...
Anyway, how this Zen chip performs will come down to how well their new motherboards perform for it. I was NOT impressed with the current AM2/3+ boards with their limited bandwidths and PCI-e 2.0 slots...
I skipped to whole FuryX VS TitanX storm of statements.
Zen will be interesting. But no matter what it will be an upgrade from the FX-8320 I am running.
On the PCI-e slots: Have you ever maxed one out?
server ssd's are the only drives I've seen that can max out a PCIE 3.0 connection :P 2.0 was and still is good enough for gfx cards. Only storage drives see a difference right now.
Yes, I believe so. with texture modding in BETH games its quite easy, and in GTA5 (4k). The general result is stuttering and such.
Textures normally fill up VRAM and then system RAM leading to stuttering. What kind of GPU were you running?
1 Like
That was my point. I feel like all the chipmakers do this to us, and we know they do. You really can't put up practical application as to the card's performance till its out, or at least till the review cards come out.
I mean, when I watched the presentation, I went 'aw, cool, but we will see.'
Plus marketing presentations are a lot of shill and nonsense to sell sell sell.
Back then I had a single 290x, I went to a Intel based system afterwards and the issues went away in regards to that. The problems were only present at 4k, I might have had 2x290x cards in-fact. But yeah, I dunno, perhaps it was just something else?
I will likely get a Zen chip once they come out as this Intel motherboard is pissing me off a bit (wants to boot into egrub all the time! and doesn't power down the USB ports fully sometimes resulting in issues).
The AM4 socket is said to support DDR4 and PCI-E 3.0 among other things.
As a user of an AM3+ Motherboard the upgrade is good, but I don't see how badly it is with PCI-E 2.1. It's just numbers really and we're only just getting into PCI-E 3.0 performance differences, before it wasn't a big deal, probably still isn't now for sure... Anyway I digress.
Here's an article because I can't be bothered explaining it... And then some ;)
2011 you know as much as everyone hypes it up.. is actually really old technology ... if your gonna put a chip on 2011 bin it from a xeon and laze off its extra cores.. your stripping a ton of value from it and overcharging quite a bit..
that being said AMD might just have to make a seperate enthusiast platform to beat intel. So they can control it better similar to what intel does and also charge huge amounts.. which everyone will bitch about.. Im gonna watch and laugh
Hyperthreading will be nice
How amd is gonna beat intel if they perform super closely to one another is microcode and scheduling.. intel still has a huge upper hand here but i would say getting very near intel.. and not even beating them is a feat by itself
http://www.techpowerup.com/223163/amd-socket-am4-platform-demo-motherboard-pictured
AM4 socket pictured, and THANK GOD it is ZIF PGA..
I am having some issues with this actually being true... First off, there is no 24pin power connector. But let's say it's an engineering sample or something just for testing... Now here comes my biggest reason to not believe this: 4pin EPS power and PCI slot. I guess it is just experimental testing board. Everything is bugging me about this board...
What do you think?
1 Like
It is a mobo for a prebuilt machine, so it doesn't strike me as impossible that it doesn't use a 24pin connector. Other than that, the only thing that strikes me is that it is wider than normal. Though that may again be due to it being for a prebuilt machine, a custom mobo so to speak. Or, it might be that the mobo is hard to shrink, meaning that miniITX will be hard to accomplish. Only time will tell.
Why do you think AMD is waiting for Sep to announce/release Zen? You think its just a manufacturing thing where they must wait for yields? or a software/firmware optimization/fixing issue?
It's more of a settings than manufacturing. If they want to start selling in September, they should start shipping early August. They should be ready with everything by now. It is more of a chipset settings, board partners preparations, etc. Lisa Su basically said "we start shipping to partners in 2 weeks", so in 2 weeks Gigabyte, Asus, MSI and Asrock will have the chips to set up bioses and what not. Then they need to make the motherboard SKUs, and so on... It is full SOC, so basically everything will be inside the chip. Motherboards should be extremely cheap unless we go again in RGB territory and have 300$ boards... Or even more.
PS: I am glad they are taking their time. I want some real competition. 1700$ is not even funny.
1 Like